4Journal of Neurology, Neurosurgery, and Psychiatry 1992;55:466-469

Unilateral brain damage after prolongedhemiconvulsions in the elderly associated withtheophylline administration

Hideo Mori, Toshio Mizutani, Masahiro Yoshimura, Hiroshi Yamanouchi,Hiroyuki Shimada

AbstractThe brains of 14 elderly patients who diedafter status epilepticus were examinedpathologically. Three of the 14 patientsshowed unilateral brain damage whichcorresponded to the side of the seizures,and the lesions were thought to be causedby seizures. In these three and two otherpatients, no causative lesion related to theseizures was found. In these five patients,status epilepticus occurred during the-ophylline therapy and thus the seizures inthese five patients were assumed to beinduced by theophylline. In the threepatients with unilateral brain damage, thedamage was seen in the hippocampus,amygdala and thalamus in two patients,while in the third, the whole hemispherewas damaged, including the hippocam-pus, amygdala, thalamus, basal gangliaand cerebral cortex. The distribution ofpathological changes within the thalamuswas described. It is suggested that thethalamus was primarily affected by seiz-ures, rather than by secondary degenera-tion from the cortex.

epilepticus to discover the effect of convulsiveseizures on the brain. Three of the 14 patientsshowed unilateral brain damage correspondingto the side of the seizures. The distribution ofpathological changes due to the unilateralbrain damage, especially in the thalamus, was

examined. The role of theophylline was dis-cussed as a possible cause of the hemiconvul-sions which resulted in unilateral brain dam-age.

Material and methodsThe brains were examined from 14 patientsover the age of 65 who had died after statusepilepticus. Coronal sections were examinedmacroscopically after formalin fixation.Paraffin blocks of the cerebrum including thethalamus and hippocampus, cerebellum, andbrainstem were prepared. Sections of paraffinblocks were stained with hematoxylin-eosin,Kliuver-Barrera stain and Holzer stain. Clinicalrecords of the patients were reviewed to obtaininformation on the underlying diseases, fea-tures of seizures and metabolic disturbance.

Department ofNeuropathology, TokyoMetropolitan Instituteof Gerontology, TokyoH MoriT MizutaniM YoshimuraDepartment ofNeurology, TokyoMetropolitan GeriatricMedical Centre, TokyoH YamanouchiDepartment ofPathology, TokyoMetropolitan GeriatricMedical Centre, Tokyo,JapanH ShimadaCorrespondence to:Dr Mori, Department ofNeurology, School ofMedicine, JuntendoUniversity, 2-1-1 Hongo,Bankyo-Ku, Tokyo 113,Japan.Received 15 May 1991 andin final revised form 13September 1991.Accepted 18 September1991

It has been thought that the adult brain israrely affected by seizures, whereas infants andyoung children are susceptible to brain damagedue to convulsive seizures. The age factor hastherefore been emphasised for brain damagefollowing status epilepticus. '3 The patholog-ical features of brain damage in seizure havebeen well documented. Seizure damage, how-ever, is sometimes difficult to distinguish froman anoxic change due to hypotension orrespiratory arrest, which often occurs in statusepilepticus, as both anoxia and seizures causesimilar histological changes, such as ischaemiccell change (ICC).We performed a neuropathological study on

14 elderly patients who died after status

Results

Three of the 14 patients showed unilateralbrain damage which corresponded to the sideof the seizures. In two other patients, damagewas found in both hemispheres: one patientshowed ICC in the hippocampus and partialloss of Purkinje cells in the left cerebellum. Theother patient showed widespread lesions inboth cerebral hemispheres and the cerebellum.No causative lesion was found in five patientsincluding the three patients with unilateralbrain damage. All of these five patients hadreceived theophylline and/or aminophylline,which contains 80% theophylline (table 1). Nometabolic disturbance which could evoke seiz-ures was found in these patients. In the otherpatients, brain lesions which could cause seiz-

Table 1 Clinical data ofpatients with status epilepticus who had no causative lesion, but received theophyUline therapy

Patient Theophylline or Basic Duration Survivalnumber Age Aminophylline disease Type of seizures ofSE after SE

1 81 Theophylline 600 mg/day PO CRF Lt-hemiconvulsion 2 days 4 days+ Aminophylline 250 mg/8h IV

2 85 Theophylline 300 mg /day PO Bronchial Lt-Jacksonian seizure 4 days 7 days-+Aminophylline 250 mg/8h IV asthma

3 77 Aminophylline 750 mg/day IV Pneumonia Rt-hemiconvulsion 2 days 1 month4 85 Aminophylline 300 mg/day PO COLD Generalised convulsion 2 hours 1 day

-iAminophylline 375 mg/day IV5 87 Aminophylline 600 mg/day IV Lung cancer Lt-focal seizure 2 hours 1 day

with generalisation

SE: Status epilepticus. IV: intravenous injection. PO: Per os. CRF: Chronic respiratory failure. COLD: Chronic obstructive lungdisease.

466 on A

pril 11, 2021 by guest. Protected by copyright.

http://jnnp.bmj.com

/J N

eurol Neurosurg P

sychiatry: first published as 10.1136/jnnp.55.6.466 on 1 June 1992. Dow

nloaded from

Unilateral brain damage after prolonged hemiconvulsions in the elderly associated with theophylline administration

A r

¢ -4. . S . @ ' ' CE-; :-

at

-c ;

Y

Figure 1 (a) Severe loss of neurons, small vacuoles and proliferation of astrocytes in thedorsomedial nucleus of the right thalamus. Patient 1. (Hematuaylin and eosin stain). (b)Proliferation of astrocytes in the pulvinar of the right thalamus. Patient 1. (Holzer stain).(c) Neurons showing shrunken cell bodies and darkly stained nuclei (c-i), and spheroids(c-2) in the lateral posterior nucleus of the right thalamus. Patient 1. (Hematoxylin andeosin stain); (d) Spongy changes and proliferation of astrocytes in the left insular cortex.Patient 3. (Hematoxylin and eosin stain).

ures were found: cerebral infarctions in fourpatients, trauma in three and tumours in two.Of the patients with cerebral infarctions, twohad received theophylline treatment, of whomone had a history of seizures before thetheophylline administration, but the other hadseizures only after the theophylline therapy. Inthe three patients with unilateral brain damage,lesions were found in the hippocampus, amyg-

dala and thalamus in two (patients 1 and 2),the hemisphere involved including the hippo-campus, amygdala, thalamus and cerebral cor-tex being affected in the other patient (patient3). The clinical histories and neuropathologicalfindings in these three patients with unilateralbrain damage were as follows.

Patient 1

An 81 year old man, who had suffered fromchronic respiratory failure since the age of 77and had received theophylline orally, wasadmitted to hospital for dyspnoea. Five daysafter admission, he had a right hemiconvulsionof a short duration. Two months later, hereceived an injection of aminophylline anddeveloped a left hemiconvulsion four days afterthe injection, which continued for two days.During the convulsion, his respiration stoppedand so he was intubated. During the status

epilepticus, the serum concentration of theo-phylline was high (26 mg/L) and so theophyl-line was stopped two days after the convulsion.Blood gas analysis and biochemical profilewere normal while the patient was receivingartificial ventilation. He remained in coma anddied four days after the status epilepticus.

Neuropathological findingsThe brain weighed 1510 gm and the cerebrumwas swollen. There was no ventricular dilata-tion. Discolouration of the right thalamus wasnoted. The main histopathological findingswere in the right hippocampus, amygdala andthalamus. In the right hippocampus, there wasa moderate loss and ICC of pyramidal neurons

in CAl, and a mild loss of neurons withproliferation of astrocytes in CA4, whereasonly a few neurons with ICC were found inCA2 and CA3. The right amygdala showed a

moderate loss of neurons with proliferation ofastrocytes in the corticomedial nuclear group.In the right thalamus, the dorsomedial nucleus(DM) showed a severe loss ofneurons, markedproliferation of astrocytes and fine spongychanges (figure a). In the ventral poster-omedial nucleus (VPM), a marked decrease inmyelinated fibres and a severe loss of neuronswere noted. In the lateral posterior nucleus(LP), ventral lateral nucleus (VL), pulvinar(Pu) and anterior nuclear group (AN), therewas a moderate loss of neurons and prolifera-tion of astrocytes (figure Ib), the remainingneurons showing ICC (fig ic-i). Many sphe-roids were noted in LIP (fig lc-2). The neuronsin the centromedian nucleus (CM), paraf-ascicular nucleus, reticular nucleus (RN) andmidline nuclei (Mid) were preserved, althoughthere was mild proliferation of astrocytes in theCM, and a few neurons in the midline nucleishowed ICC (table 2). There was mild pro-liferation of astrocytes in the right cerebralcortex. There was no notable change in theright caudate, right putamen, right pallidumand right subthalamus. There was a moderateloss of granule cells in the cerebellum and a

partial loss of Purkinje cells in the left cere-bellar hemisphere. The left cerebral hemi-sphere and the brainstem were unremarkable.

Patient 2An 85 year old woman was admitted tohospital for fever and stridor. She had beenrepeatedly admitted for bronchial asthma sinceage 79 and had received theophylline orally.After admission, aminophylline was admin-istered intravenously. On the third day ofhospitalisation, she developed Jacksonian seiz-ures, which started from the left upper limb

Table 2 Distribution of the seizure damage in the thalamic nuclei in patients of the unilateral brain damage

Patientnumber side AN DM Mid CM LD LP Pu VA VL VPL VPM RN

1 Rt ++ +++ i i ++ ++ ++ + ++ + +++ -2 Rt NA +++ ± i +++ ++ NA NA NA ++ ++ -3 Lt +++ +++ i i ++ ++ +++ + +++ ++ +++ -

AN: Anterior nuclear group. DM: Dorsomedial nucleus. Mid: Midline nuclei. CM: Centromedian nucleus. LD: Lateral dorsalnucleus. LP: Lateral posterior nucleus. Pu: Pulvinar. VA: Ventral anterior nucleus. VL: Ventral lateral nucleus. VPL: Ventralposterolateral nucleus. VPM: Ventral posteromedial nucleus. RN: Reticular nucleus.-: none. ±: slight. +: mild. ++: moderate. +++: severe. NA: not available.

467 on A

pril 11, 2021 by guest. Protected by copyright.

http://jnnp.bmj.com

/J N

eurol Neurosurg P

sychiatry: first published as 10.1136/jnnp.55.6.466 on 1 June 1992. Dow

nloaded from

Mon, Mizutani, Yoshimura, Yamanouchi, Shimada

and spread to the whole body. The convulsionsoccurred repeatedly and lasted for four days.Blood gas analysis showed the following: pH7 4; Po2, 89 mmHg; and Pco2, 27 mmHg.Biochemical tests were normal except bloodglucose, 179 mg/dL. She remained in comaand died seven days after the status epilepti-cus.

Neuropathological findingsThe brain weighed 1050 gm. Small infarctionswere found in the white matter of the leftfrontal lobe and the right putamen. Theventricle was not enlarged. The main histo-logical changes were present in the righthippocampus, amygdala and thalamus. Manypyramidal neurons in the right hippocampusshowed ICC. In the right amygdala, there wassevere neuronal loss with proliferation of astro-cytes, particularly in the corticomedial nucleargroup. In the right thalamus, there was a totalloss of neurons with proliferation of astrocytesin DM and the lateral dorsal nucleus (LD).The neuronal loss was less marked in LP andthe ventral posterolateral nucleus (VPL).Spheroids and a few neurons with ICC werefound in LP. Neurons in CM were spared, butthere was mild proliferation of astrocytes (table2). In the right cerebral cortex, there was mildproliferation of astrocytes. In the right caudate,there was mild proliferation of astrocytes, butin the right putamen, right pallidum and rightsubthalamus, no obvious change was observed.In the left thalamus, a mild loss of neuronswith moderate proliferation of astrocytes wasfound in LD and DM. In the left cerebralcortex, left putamen, left pallidum, left sub-thalamus, cerebellum and brainstem, there wasno notable lesion.

Patient 3A 77 year old woman was admitted with a feverand cough. She had received injections ofaminophylline for the previous 10 days. Shehad a history of surgery for a left cerebello-pontine angle tumour at the age of 73. Right

level

hemiconvulsions occurred after admission andcontinued for two days. Blood gas analysisshowed the following: pH, 7-2; Po2,57-8 mmHg; and Pco2, 70-8 mmHg. Otherbiochemical tests were normal. She receivedartificial ventilation for one week. After theconvulsions, she showed no response and diedone month later.

Neuropathological findingsThe brain weighed 860 gm. The width of theleft cerebral cortex was decreased. The thirdventricle was enlarged. The left cerebral cortexwas entirely damaged. The severity of thedamage in the cortex varied (figure 2). Theleast affected area showed a neuronal loss withmild proliferation of astrocytes in the thirdlayer. The severely affected area showed a totalneuronal loss in all layers and spongy changewith marked proliferation of protoplasmicastrocytes. The left parahippocampal and leftoccipitotemporal gyri were relatively spared.The left insular cortex was severely damaged(figure id). In the left thalamus, AN, DM, VLand Pu showed a total neuronal loss andmarked proliferation of astrocytes. A focus ofmyelin pallor and depletion of neurons werenoted in VPM. In LP and VPL, only a fewneurons remained and astrocytes proliferatedmoderately (table 2). In the left hippocampus,there was an almost total loss of neurons withproliferation of astrocytes, except in CA2. Inthe left amygdala, the proliferation of astro-cytes was marked, but the loss of neurons wasmoderate. In the left putamen and left caudate,there was mild proliferation of astrocytes. Inthe right cerebral hemisphere, the cerebellumand the brainstem, there was no obviouschange.

DiscussionThree of the 14 patients showed unilateralbrain damage corresponding to the side ofseizures. Anoxic changes in global ischaemia(due to cardiac arrest or hypotension) orhypoxia usually occurs in both hemispheres,whereas some cases show lesions in only onehemisphere, especially in the arterial boundaryzone. In the three patients reported here, theanoxic change was restricted to the focal side ofseizures and the unilateral lesions were con-sidered to be a result of the seizures. Thepatterns of the anoxic changes produced byglobal ischaemia, hypoglycaemia, and seizureare similar, but a slight difference was recog-nised: the cerebellum was less frequently affec-ted in status epilepticus than after cardiacarrest.4 In the three patients reported, thecerebellum either escaped damage (patients 2and 3) or was only mildly affected (patient 1).These findings also support the idea that thelesions seen in the three patients were causedby seizures. Seizures have been assumed to beone ofthe causes ofhemiatrophy in children.56Recently, Soffer et al7 reported hemisphericbrain damage after unilateral status epilepticusin a young adult, which involved the cerebralcortex, thalamus and hippocampus on thefocal side of the convulsions. The distribution

Severe 1 Mild

_ Moderate Slight

Figure 2 Distribution of the cortical lesions in patient 3. (left: frontal lobe, middle:of the thalamus, right occipital lobe).

468 on A

pril 11, 2021 by guest. Protected by copyright.

http://jnnp.bmj.com

/J N

eurol Neurosurg P

sychiatry: first published as 10.1136/jnnp.55.6.466 on 1 June 1992. Dow

nloaded from

Unilateral brain damage after prolonged hemiconvulsions in the elderly associated with theophylline administration

of the seizure damage in this patient is similarto that in patient 3.

In our three patients no causative lesion forthe seizures was found. The three patients andtwo others (patients 4 and 5 in table 1), whoalso showed no causative lesion for the seiz-ures, had seizures during the administration oftheophylline. Seizures have been reported to bea complication of theophylline administration.Theophylline-induced seizures are often focalor unilateral, especially in adults, and arerefractory to anticonvulsant.8'0 Recent obser-vations have shown that theophylline-inducedseizures are caused by adenosine receptorantagonism." There are two forms of the-ophylline toxicity. One is acute toxicity due toa single large injection and the other is thechronic toxicity that develops with long-termtherapy.2 13" e chronic toxicity of theophyl-line was assumed to be the cause ofthe seizuresin these five patients. The serum concentrationof theophylline was assessed in only one of thepatients in this series. An earlier study revealedseizures occurring with high serum theophyl-line concentrations,9 but a recent study sug-gested that there is no relationship between theseverity of the adverse effects and the serumconcentration of theophylline.'2 Adverseeffects could occur at lower serum concentra-tions with chronic intoxication.'3 Nakada etal'0 reported that seizures could occur attherapeutic levels, especially in the elderly withlow serum albumin levels. Of the five necropsypatients reported by Yarnell and Chu8 andNakada et al,o1 only one showed a neuronalloss in the hippocampus and cerebellum,which might have been the result of seizures.On the other hand, Culberson et al "1 describedtwo patients who showed an anoxic change inthe unilateral cerebral cortex, thalamus andhippocampus corresponding to the focal sideof seizures in five necropsy patients. Noetzel"reported a young child who showed unilateralcerebral hemiatrophy on CT scanning longafter theophylline-induced seizures. The uni-lateral brain damage in the patients of Culber-son et al and the hemiatrophy in the patient ofNoetzel can be interpreted to be consequencesof seizure damage.The thalamus was uniformly involved in our

patients. Scholz5 first pointed out that thethalamus is frequently as damaged as thehippocampus and cerebellum following seiz-ures. In our patients, the thalamic lesion wasextensive and DM was always severely dam-aged. On the other hand, the nuclei withsubcortical or diffuse cortical projections, suchas CM and RN, tended to be spared (table 2).Meyer et al 6 regarded the thalamic lesions as

retrograde degeneration from the cortex, asnon-specific nuclei such as CM were spared.Conversely, Tan and Urich6 presented the ideaof the thalamic lesions representing ictal braindamage. The thalamic lesions in patients 1 and

2 in our series demonstrated that the thalamuswas damaged by the seizures themselves, not

by secondary degeneration from the cortex, asthe corresponding cortices were little affectedin these patients. ICC of the neurons in thethalamus also supported this concept.

It has been shown that seizure activity itselfevoked brain damage in experimental animalsand thalamic lesions could result from seizuresinduced by systemic administration or cerebralinjection of epileptogenic agents.7'-9 Recentexperimental studies on animals suggested thatthe excessive release of endogenous excitatoryamino acids, particularly glutamate, is impli-cated in ictal damage of the hippocampus andthe same mechanism is also postulated forthalamic lesions.'8 It is unknown whether ictalbrain damage in humans can be explained bythe same mechanism. Corticothalamic fibresare assumed to be glutamatergic, so the tha-lamic lesions might be explained by this theory.Whether or not the pattern of thalamic lesionscorresponds to the distribution of glutamatereceptors remains unknown.

1 ScholzW. The contribution of patho-anatomical research tothe problem of epilepsy. Epilepsia 1959;1:36-55.

2 Oxbury JM, Whitty CWM. Causes and consequences ofstatus epilepticus in adults. A study of 86 cases. Brain197 1;94:733-44.

3 Corsellis JAN, Bruton CJ. Neuropathology of status epi-lepticus in humans. Adv Neurol 1983;34:129-39.

4 Ng T, Graham DI, Adams JH, Ford I. Changes in thehippocampus and the cerebellum resulting from hypoxicinsults: frequency and distribution. Acta Neuropathol(Bert) 1989;78:438-43

5 Scholz W. Die Krampfschadigungen des Gehirns, Berlin:Springer, 1951.

6 Tan N, Urich H. Postictal cerebral hemiatrophy: with acontribution of the problem of crossed cerebellar atrophy.Acta Neuropathol (Berl) 1984;62:332-9.

7 Soffer D, Melamed E, Assaf Y, Cotev S. Hemispheric braindamage in unilateral status epilepticus. Ann Neurol1986;20:737-40.

8 Yarnell PR, Chu NS. Focal seizures and aminophylline.Neurology 1975;25:819-22.

9 Zwillich CW, Sutton FD Jr, NeffTA, Cohn WM, MatthayRA, Weinberger MM. Theophylline-induced seizures inadults. Correlation with serum concentrations. Ann InternMed 1975;82:784-7

10 Nakada T, Kwee IL, Lerner AM, Remler MP. Theophyllineinduced seizures: clinical and pathophysiologic aspects.West JrMed 1983;138:371-4.

11 Dragunow M. Adenosine receptor antagonism accounts forthe seizure-prolonging effects of aminophylline. Pharna-col Biochem Behav 1990;36:751-5.

12 Bertino JS Jr, Walker JW Jr. Reassessment of theophyllinetoxicity. Serum concentrations, clinical course, and treat-ment. Arch Intern Med 1987;147:757-60.

13 Olson KR, Benowitz NL, Woo OF, Pond SM. Theophyllineoverdose: acute single ingestion versus chronic repeatedovermedication. Am J Emerg Med 1985;3:386-94.

14 Culberson CG, Langston JW, Herrick M. Aminophyllineencephalopathy: a clinical, electroencephalographic, andneuropathological analysis. Trans Am Neurol Assoc 1979;104:224-6.

15 Noetzel MJ. Theophylline neurotoxicity resulting in sig-nificant unilateral brain-damage. Dev Med Child Neurol1985;27:242-5.

16 Meyer A, Beck E, Shepherd M. Unusually severe lesions inthe brain following status epilepticus. J Neurol NeurosurgPsychiary 1955;18:24-33.

17 Collins RC, Olney JW. Focal cortical seizures cause distantthalamic lesions. Science 1982;218:177-9.

18 Ben-AriY,Tremblay E, Ottersen OP, Meldrum BS.The roleof epileptic activity in hippocampal and 'remote' cerebrallesions induced by kainic acid. Brain Res 1980;191:79-97.

19 Ingvar M, Morgan PF, Auer RN. The nature and timing ofexcitotoxic neuronal necrosis in the cerebral cortex,hippocampus and thalamus due to flurothyl-inducedstatus epilepticus. Acta Neuropathol (Bert) 1988;75:362-9.

469 on A

pril 11, 2021 by guest. Protected by copyright.

http://jnnp.bmj.com

/J N

eurol Neurosurg P

sychiatry: first published as 10.1136/jnnp.55.6.466 on 1 June 1992. Dow

nloaded from